Cutlery/calibre
1
0
Fork 0
mirror of https://github.com/kovidgoyal/calibre.git synced 2025-07-09 03:04:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

Strip comments from within <style> tags as they can sometimes confuse BeautifulSoup

Browse Source
This commit is contained in:
Kovid Goyal 2008-01-03 21:47:17 +00:00
parent 7ea58726b4
commit ad1ec0f474
1 changed files with 16 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
src/libprs500/ebooks/lrf/html/convert_from.py
View File

@ -65,6 +65,19 @@ def munge_paths(basepath, url):
path = os.path.join(os.path.dirname(basepath), path) path = os.path.join(os.path.dirname(basepath), path)
return os.path.normpath(path), fragment return os.path.normpath(path), fragment
def strip_style_comments(match):
src = match.group()
while True:
lindex = src.find('/*')
if lindex < 0:
break
rindex = src.find('*/', lindex)
if rindex < 0:
src = src[:lindex]
break
src = src[:lindex] + src[rindex+2:]
return src
class HTMLConverter(object): class HTMLConverter(object):
SELECTOR_PAT = re.compile(r"([A-Za-z0-9\-\_\:\.]+[A-Za-z0-9\-\_\:\.\s\,]*)\s*\{([^\}]*)\}") SELECTOR_PAT = re.compile(r"([A-Za-z0-9\-\_\:\.]+[A-Za-z0-9\-\_\:\.\s\,]*)\s*\{([^\}]*)\}")
PAGE_BREAK_PAT = re.compile(r'page-break-(?:after|before)\s*:\s*(\w+)', re.IGNORECASE) PAGE_BREAK_PAT = re.compile(r'page-break-(?:after|before)\s*:\s*(\w+)', re.IGNORECASE)
@ -87,6 +100,9 @@ class HTMLConverter(object):
# Replace entities # Replace entities
(re.compile(ur'&(\S+?);'), partial(entity_to_unicode, (re.compile(ur'&(\S+?);'), partial(entity_to_unicode,
exceptions=['lt', 'gt', 'amp'])), exceptions=['lt', 'gt', 'amp'])),
# Remove comments from within style tags as they can mess up BeatifulSoup
(re.compile(r'(<style.*?</style>)', re.IGNORECASE|re.DOTALL),
strip_style_comments),
] ]
# Fix Baen markup # Fix Baen markup
BAEN = [ BAEN = [